Random Forest for IoT Smart the home Attack and Anomaly Detection

Abstract

The Internet of Things' growth (IoT) has also led to an increase in damaging attacks that seriously jeopardize unprotected IoT equipment. IoT technologies therefore raise a number of security and privacy concerns.  Anomalies in IoT systems may enable an attacker to infiltrate a system, often resulting in unforeseen interruptions or sensor malfunctions. Consequently, it is indispensable to monitor the unforeseen occurrences resulting from sensor inputs.   This study evaluates the efficacy of an ensemble mechanism knowledge model against a conventional learning method for detecting attacks and anomalies inside an IoT smart home environment, using a categorical dataset of mainSimulationAccessTraces traffic data.   To achieve optimal anomaly detection outcomes, we use the Random Forest methodology to consolidate the most effective models subsequent to training each model using the training dataset.  This paper elucidates many metrics and evaluation models while also providing a complete explanation of the training strategy.   We also compare our findings to those of comparable model versions.   Finally, we discuss difficulties and future projects.